SPERMATOGENESIS OF PANDARUS SINUATUS. I I 5 



to grow large. The chromosomes in the early prophase (Fig. 5) 

 are thread-like and do not show a longitudinal split as is the case 

 in the primary oocytes of this species. It may be that the long- 

 itudinal split is present but cannot be seen on account of the 

 smaller size of the chromosomes, or the splitting may occur after 

 the synapsis, at which time a division is indicated by a constric- 

 tion at right angles to the plane dividing the chromosomes of 

 each pair (see below). Another interpretation is that in the 

 primary oocytes the " split " represents the division between ad- 

 jacent chromosomes and is therefore after the synapsis. The 

 chromosomes are at first sixteen in number, but as they become 

 denser some appear to be joined end to end (Fig. 6), and by this 

 time the nucleoli have entirely dissolved. The chromosomes 

 collect together in a dense mass so that only their ends sticking 

 out can be distinguished separately (Fig. 7), and after the ele- 

 ments of this mass separate they are seen to be eight double 

 chromosomes united end to end (Fig. 8). The chromosomes 

 now shorten, at the same time becoming thicker (Fig. 9), and 

 soon a second constriction transforms each double chromosome 

 into a tetrad (Fig. 10). Each tetrad continues to shorten until 

 the width is as great as the length (Fig. 1 1). The nuclear wall 

 dissolves and the spindle is now formed (Figs. 12, 13, 14). In 

 the equatorial plate seven tetrads are arranged in a circle and the 

 eighth lies in the center. It is impossible to observe whether the 

 division is reducing or not, owing to the shape of the tetrads. 



The second spermatocytic division follows immediately after 

 the first. Each of the eight diads is divided equally between the 

 two daughter cells (Figs. 15, 16). In the spermatids thus 

 formed the chromosomes swell up and fuse to form nuclei, and 

 the spermatid (Fig. 18) resembles the primary spermatocyte save 

 for the reduction in the size and the absence of nucleoli. 



By the methods used no distinction could be made out between 

 the spermatids when first formed, but they develop into structures 

 that show as great differences as exist between the products of 

 the testes of any species of animals that has come to my attention. 

 Many spermatids degenerate and appear to be absorbed as food 

 by those remaining. Some of them elongate (Figs. 19 and 20) 

 and begin to develop into spermatozoa. Whereas the spermatids 



